South Africa has on of the highest rates of HIV infection in the world, and drug use among its youth has been on the increase during the 1990s. South Africa's socio-political change from apartheid to democracy is considered to be a main contributor to these changes. This proposal is for a collaborative study of risk and protective factors for HIV risk behaviors and drug use of adolescents in South Africa. The main HIV risk behaviors that will be examined are engagement in sexual interactions with multiple partners, and without a condom. The use of drugs refers to both legal and illegal drugs. The proposed study will examine among South African adolescents: (1) the relationship between HIV risk behaviors and drug use; (2) the family, personality, peer, demographic, ecological and cultural ('family interactional') risk and protective factors that relate to HIV risk behaviors and drug use; (3) the role of social cognitive factors (of the theory of planned behavior) within the family interactional model in accounting for HIV risk behaviors and drug use; and (4) the shared and unshared risk factors for HIV risk behaviors and drug use. The study will be conducted in two comparison cities (Durban and Cape Town) that are located in the provinces with respectively, the highest and the lowest rate of HIV infection. Within a cross-sectional study, 600 male and female adolescents (aged 12-17 years) will be recruited in each site to be interviewed in a community survey. They will be asked about their drug use and HIV risk behaviors, personality attributes, familial relations and drug involvement, relations with peers, cultural orientation and their neighborhood environment. They will also be asked about their intentions, attitudes, subjective norms, and perceptions of control (based on the theory of planned behavior) in relation to the target behaviors, and other social cognitive factors. The final questions of the interviews will be designed following focus groups and pre-testing of the instrument with adolescents from the target population. Measures for the study will be developed with the use of factor analytic and reliability techniques. The data will be analyzed using Pearson's product moment correlation techniques, multiple regression analyses, hierarchical multiple regression analyses, structural equation modelling and net regression analyses. In assessing and comparing the risk and protective factors for each behavior the study should shed light on potentially beneficial approaches for understanding and developing interventions for their prevention.